Leucine Aminopeptidase Is Not Essential for Trimming
Peptides in the Cytosol or Generating Epitopes for MHC Class
I Antigen Presentation1
Charles F. Towne,* Ian A. York,* Joost Neijssen,†Margaret L. Karow,‡Andrew J. Murphy,‡
David M. Valenzuela,‡George D. Yancopoulos,‡Jacques J. Neefjes,†and Kenneth L. Rock2*
To detect viral infections and tumors, CD8?T lymphocytes monitor cells for the presence of antigenic peptides bound to MHC
class I molecules. The majority of MHC class I-presented peptides are generated from the cleavage of cellular and viral proteins
by the ubiquitin-proteasome pathway. Many of the oligopeptides produced by this process are too long to stably bind to MHC class
I molecules and require further trimming for presentation. Leucine aminopeptidase (LAP) is an IFN-inducible cytosolic amino-
peptidase that can trim precursor peptides to mature epitopes and has been thought to play an important role in Ag presentation.
To examine the role of LAP in generating MHC class I peptides in vivo, we generated LAP-deficient mice and LAP-deficient cell
lines. These mutant mice and cells are viable and grow normally. The trimming of peptides in LAP-deficient cells is not reduced
under basal conditions or after stimulation with IFN. Similarly, there is no reduction in presentation of peptides from precursor
or full-length Ag constructs or in the overall supply of peptides from cellular proteins to MHC class I molecules even after
stimulation with IFN. After viral infection, LAP-deficient mice generate normal CTL responses to seven epitopes from three
different viruses. These data demonstrate that LAP is not an essential enzyme for generating most MHC class I-presented peptides
and reveal redundancy in the function of cellular aminopeptidases. The Journal of Immunology, 2005, 175: 6605–6614.
expressed on most nucleated cells, and consists of a genetically
polymorphic H chain, an L chain (?2-microglobulin), and a small
peptide, usually 8–10 aa long. This peptide provides specificity for
recognition by the TCR and is produced by proteolysis of cellular
proteins, predominately by the proteasome (1, 2).
Purified proteasomes, when incubated with full-length proteins
such as chicken OVA, casein, or insulin-like growth factor, gen-
erate many peptides, ranging in length from three to 22 aa (3–5).
Only ?15% of the peptides produced by digestion of these pro-
teins are of optimal binding length for MHC class I ligands, but an
additional 15–25% of the peptides are longer than 10 residues and
could therefore be further processed to produce MHC class I-bind-
ing peptides (3, 4, 6). In vivo, proteasomes also generate longer
peptides that are trimmed and presented (7–10). In fact, it was
recently suggested that most peptides produced by proteasomes in
vivo are longer than 15 residues (11). In any case, the majority of
peptides generated by the proteasome are destroyed by cytosolic
peptidases before they encounter the TAP transporter (11, 12).
tytotoxic T lymphocytes recognize virus-infected or oth-
erwise abnormal cells through TCR interactions with
MHC class I molecules (MHC class I). MHC class I is
Treating cells with proteasome inhibitors almost completely
prevents presentation of peptides not only from full-length protein
precursors, but also from peptide precursors that are extended even
by one amino acid at the C terminus (13, 14). Therefore, the pro-
teasome appears to be required for generation of the C-terminal
amino acid of most MHC class I-binding peptides. However, pep-
tides derived from precursors that are extended by 25 aa at the N
terminus can be efficiently presented on MHC class I even in the
presence of proteasome inhibitors (14), but only if the N terminus
is unblocked (13), implicating aminopeptidases in this form of
Some precursor peptides are trimmed in the endoplasmic retic-
ulum (ER)3by the aminopeptidase ERAP1 after transport to the
ER by TAP (7, 9, 10). In cells lacking ERAP1, N-extended pre-
cursors that are transported into the ER fail to be trimmed and
presented (10). However, in the absence of ERAP1, there is still
trimming and presentation of a significant fraction of precursor
peptides by aminopeptidases in the cytosol (10). There are several
aminopeptidases in the cytosol that can potentially trim peptides.
One of the first aminopeptidases to be implicated in Ag pro-
cessing was leucine aminopeptidase (LAP), which was identified
as a major activity in cytosolic extracts from HeLa cells that was
capable of trimming N-extended precursors of SIINFEKL, an an-
tigenic peptide from OVA, to the mature presented epitope (15).
Another indication that this enzyme might play an important role
in MHC class I peptide processing is that transcription of the LAP
gene is strongly up-regulated by IFN-?, a cytokine that, among its
many proinflammatory effects, up-regulates many of the essential
*Department of Pathology, University of Massachusetts Medical School, Worcester,
MA 01655;†Division of Tumor Biology, The Netherlands Cancer Institute, Amster-
dam, The Netherlands; and‡Regeneron Pharmaceuticals, Tarrytown, NY 10591
Received for publication May 23, 2005. Accepted for publication September 8, 2005.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1This work was supported by a National Institutes of Health grant (to K.L.R.) and
National Institutes of Health Training Grant AI07349 (to C.F.T.). Core resources
supported by Diabetes Endocrinology Research Grant DK42520 were also used.
2Address correspondence and reprint requests to Dr. Kenneth L. Rock, Department of
Pathology, University of Massachusetts Medical Center, Room S2-109, 55 Lake Avenue
North, Worcester, MA 01655-0125. E-mail address: firstname.lastname@example.org
3Abbreviations used in this paper: ER, endoplasmic reticulum; BH, bleomycin hy-
drolase; ERAP, ER aminopeptidase; ES, embryonic stem; KD, knockdown; KO,
knockout; LAP, leucine aminopeptidase; LCMV, lymphocytic choriomeningitis vi-
rus; MEF, mouse embryonic fibroblasts; poly I:C, polyinositic-polycytidylic acid;
PSA, puromycin-sensitive aminopeptidase; shRNA, short-hairpin RNA; siRNA,
small interfering RNA; TOP, thimet oligopeptidase; Vac, vaccinia; VSV, vesicular
stomatic virus; WT, wild type.
The Journal of Immunology
Copyright © 2005 by The American Association of Immunologists, Inc.0022-1767/05/$02.00
components of Ag presentation (15). Moreover, overexpressing
LAP in cells led to more rapid trimming of peptides (11, 16). For
these reasons it was thought that LAP was an important peptidase
in generating MHC class I-presented peptides.
In addition to LAP, mammalian cells contain many other pep-
tidases in the cytosol, and several of these have been proposed to
play a role in Ag presentation in certain situations (1). Puromycin-
sensitive aminopeptidase (PSA) and bleomycin hydrolase (BH) are
cytosolic aminopeptidases that were shown to trim a vesicular sto-
matic virus (VSV) precursor peptide in cell extracts (17). Treat-
ment of cells with inhibitors that block PSA and BH activity,
among other things, reduced the presentation of the VSV epitope
(17). Tripeptidyl peptidase II has exo- and endopeptidase activities
and may play a unique role in the initial trimming of very long
peptides (?16 residues) in the cytosol (8). A major unresolved
question is whether the various aminopeptidases have unique roles
in protein catabolism and Ag presentation or whether there is sig-
nificant functional redundancy.
In addition to generating mature MHC class I-binding peptides
from precursors, peptidases have the potential to destroy epitopes
by trimming them to a size that is too short to bind MHC class I.
ERAP1 and the endopeptidase thimet oligopeptidase (TOP) have
been shown to destroy peptides in cells and limit Ag presentation
(9, 10, 18). In addition, PSA (19) and LAP (11, 20) have been
shown to have the potential to destroy antigenic peptides, although
the extent to which they limit Ag presentation in vivo, if any, is not
In this study we demonstrate that mice lacking LAP and human
cells in which LAP levels are reduced with small interfering RNA
(siRNA) do not exhibit any detectable differences in Ag presenta-
tion. These results demonstrate that LAP does not play an essential
role in the generation or destruction of many antigenic peptides
and indicate that there is functional redundancy among the cyto-
Materials and Methods
Plasmids, primers, and PCR
The plasmids used to make the LAP-KD and control cell lines were gen-
erated by designing short-hairpin RNA (shRNA) inserts using RNAiOli-
goRetriever software and were based on the National Center for Biotech-
nology Information sequence for human LAP, NM_015907. Synthetic
oligos (IDT) were inserted into BseRI and BamHI digested pSHAG-1 plas-
mid (21) (gift of Dr. G. Hannon, CSH Laboratories, Cold Spring Harbor,
NY). The oligos used were: for LAP, GAATGATCCCATTGCCTGTTC
ATTGAGGAACAGGCAATGGGATCATTCcg; and for the control cell line,
CTTCTTGGTGTCAGGATAGtttttt, and gatcaaaaaaCTATCCTGACACCA
AGAAcg. Kanamycin-resistant bacterial clones were selected, and the
plasmids were sequenced. pSHAG?shRNA plasmids were digested with
PvuII, and the fragment including the U6 promoter and shRNA sequence were
subcloned into pTracer-CMV2 (Invitrogen Life Technologies), which had
been digested with NruI and EcoRV, removing the CMV promoter.
Primers used for identification of shRNA knockdown (KD) clones and
for real-time PCR were 5?-GCACGCCAATTGATGGAG-3? and 5?-
GGTCTGATATGGACCTCG-3?. ?-Actin was used as a control for nor-
malization in real-time PCR assays. The sequences of the primers used
were 5?-CGAGGCCCAGAGCAAGAGAG-3? and 5?-CGGTTGGCCT
The SIINFEKL mini (MSIINFEKL), N5?SIINFEKL (LEQLESIIN
INFEKL), and full-length OVA genes were all subcloned from other vec-
tors into pTracer-CMV2 (Invitrogen Life Technologies), a plasmid
containing a GFP/zeocin resistance fusion protein, by restriction digest and
ligation. The plasmids were then sequenced to confirm correct sequences
and reading frames.
To express mature SIINFEKL with no N-terminal residues, we con-
structed pUG-SIINFEKL. This plasmid consists of ubiquitin with SIIN
FEKL fused to the C terminus; C-terminal ubiquitin hydrolases efficiently
release peptides thus fused to ubiquitin (23). An internal ribosome entry
site downstream of the ubiquitin-SIINFEKL fusion was followed by GFP.
mRNA was isolated from 106cells using an RNeasy kit (Qiagen). cDNA
was synthesized using SuperScript II enzyme (Invitrogen Life Technolo-
gies), and real-time PCR was performed on an iCycler machine (Bio-Rad)
with SYBR Green buffer (Applied Biosystems).
To generate the upstream and downstream homologous sequences for
ligation to the 5? and 3? ends of the knockout reporter gene, two sets of
primers were used. For the upstream sequence, the upstream forward
primer was 5?-AGACCCTAGAAAGGACGACGG-3, and the upstream re-
verse primer was 5?-GGCCCTGTGACTGGCTACTC-3?. For the down-
stream sequence, the downstream forward primer was 5?-TGGTGC
CATCTTTCTCAGGAC-3?, and the downstream reverse primer was 5?-
GTGGTCACCTTGGTCTGCAAG-3?. For screening of embryonic stem
(ES) cells that contained a knockout allele, TaqMan primers were used.
The forward sequence was 5?-AGGATTGTCCCAAAGCCTGCTACGCT
3?, and the reverse sequence was 5?-TGGTGTTCAGTGATGGAG
GTCTAGCATGCA-3?. Each primer sequence spanned the point of recom-
bination between the reporter gene and the endogenous genomic sequence.
A three-primer PCR protocol was used to screen for LAP-deficient
mice. LAP-WT(R1) 5?-CAGATATGGCTGATTCTAGC-3? lies down-
stream of the knockout (KO) insert in the genomic sequence and therefore
amplifies both the wild-type (WT) and KO alleles. LAP-KO(F4) 5?-GC
CTGAAGAACGAGATCAGC-3? lies within the KO allele and amplifies
only the KO allele. LAP-WT(F1) 5?-GCACACTTAGACATAGCAG-3?
lies within the WT allele and amplifies only the WT allele.
Generation of LAP?/?mice
The mouse LAP gene (gene identification no. 66988) was deleted using
VelociGene technology (24). Briefly, a large targeting vector (BACvec)
was constructed by bacterial homologous recombination in which the
19.2-kb LAP gene was replaced by a lacZ-neo cassette. An 129 ?
C57BL/6 F1ES cell line was electroporated with the BACvec and selected
for G418 resistance. Drug-resistant clones were screened for loss of one
copy of the LAP gene by quantitative PCR using probes at either end of the
deletion. Two independent targeted ES cell clones were microinjected into
C57BL/6 blastocysts, which were implanted in C57BL/6 females to gen-
erate chimeras. Chimeras were bred back to BL/6 to generate F1hetero-
zygote mice. Both ES cell clones produced healthy-appearing knockout
mice in proper Mendelian proportions. A line derived from one of these
clones was used for subsequent experiments.
Poly I:C (poly I:C) treatment of mice
Mice were injected i.p. with 200 ?g of poly I:C (Amersham Biosciences)
in a total volume of 200 ?l of PBS (Invitrogen Life Technologies). Spleens
from the mice were harvested after 24 h and then stained for flow cyto-
Virus infection of mice
Mice were injected i.p. with 5 ? 104PFU/mouse of lymphocytic chorio-
meningitis virus (LCMV) Armstrong (a gift from Dr. R. Welsh, University
of Massachusetts Medical School, Worcester, MA) or with 5 ? 106PFU/
mouse of recombinant vaccinia (Vac; provided by Drs. J. Yewdell and J.
Bennink, National Institutes of Health, Bethesda, MD), containing chicken
OVA (25). Mice were infected i.v. with 5 ? 106PFU/mouse of VSV (a gift
from Dr. R. Welsh). Nine days (LCMV) or 7 days (Vac-OVA and VSV)
after infection, splenocytes were harvested and incubated for 5 h with the
appropriate peptide (5 ?M for LCMV and Vac-OVA; 2 ?M for VSV) or
with anti-CD3? (BD Biosciences) in the presence of GolgiPlug (BD Bio-
sciences) and rIL-2 (BD Biosciences). Cells were then stained for CD8,
CD44, and intracellular levels of IFN-? using commercial Abs (BD Bio-
sciences) and were analyzed by flow cytometry.
Construction of stable shRNA cells
HeLa-Kbcells (10) were transfected with the shRNA?GFP/zeocin resis-
tance-expressing plasmids using HeLa Monster (Mirus). Forty-eight hours
after transfection, the cells were serially diluted in DMEM containing 100
?g/ml zeocin (Invitrogen Life Technologies) to select for plasmid trans-
fectants and seeded in 96-well plates. After 14 days of incubation in se-
lection medium, clones were isolated and tested for the expression of LAP
mRNA by RT-PCR and for GFP expression by FACS. Clones used in these
studies were selected based on comparable expression of GFP.
6606ROLE OF LEUCINE AMINOPEPTIDASE IN Ag PRESENTATION
Cells and tissue culture
The shRNA KD cell lines were incubated at 37°C and 10% CO2in
DMEM/10% FCS and 100 ?g/ml zeocin to select for cells expressing the
shRNA and in 100 ?g/ml G-418 (Invitrogen Life Technologies) to select
for cells expressing H-2Kb. Mouse embryonic fibroblasts (MEFs) were
generated from 12- to 14-day embryos and cultured at 37°C in 10% CO2
with DMEM/20% FCS. Cells were cultured in flasks, and transfections
were performed in six-well plates (Corning-Costar). During transfection
periods, cells were cultured in DMEM/10%FCS.
For incubations in IFN-?, MEFs or shRNA KD cells were transfected
first with the indicated construct. Then 8 h later, the transfection medium
was removed, and DMEM, FCS, and IFN-? were added to each well. The
IFN-? concentrations used were 250 U/ml recombinant human IFN-? for
HeLa cells (Biogen) or 50 U/ml recombinant murine IFN-? for MEFs (BD
Biosciences). The cells were incubated in IFN-? until they were analyzed.
Abs, immunoblotting, and flow cytometry
The mAb 25.D1.16 (anti-Kb?SIINFEKL) (26), AF6-88.5 (anti-Kb) (27),
Y3 (anti-Kb) (28), M1/42 (anti-H2) (29), or H36.4.5 (anti-influenza hem-
agglutinin; gift from W. Gerhard, The Wistar Institute, University of Penn-
sylvania, Philadelphia, PA) were used as primary Abs in staining HeLa
cells and MEFs for flow cytometry. After incubation in one of the primary
Abs, the cells were washed with PBS and stained with donkey anti-mouse
(or donkey anti-rat) F(ab?)2conjugated to Cy5 (Jackson ImmunoResearch
Laboratories). For staining cells isolated from spleen, AF6-88.5 (H-2Kb)
and KH95 (H-2Db) Abs conjugated to a fluorophore were used according
to the manufacturer’s directions (BD Biosciences). The cells were then
analyzed by flow cytometry on a FACSCalibur apparatus (BD Biosciences)
with FlowJo software (TreeStar).
For immunoblotting HeLa cells, cells were lysed in a 1-ml Dounce
homogenizer (Kontes) on ice in Dounce buffer (10 mM Tris-HCl and 0.5
mM MgCl2, pH 7.6). Complete EDTA-free protease inhibitor mixture
minitablets (Roche) were added according to the manufacturer’s instruc-
tions. After 50 strokes in the Dounce homogenizer, tonicity buffer was
added (10 mM Tris-HCl, 0.5 mM MgCl2, 0.6 M NaCl, and complete mini-
tablets) to the lysate to stabilize the nuclei. After a 5-min spin in a mini-
centrifuge at 4°C at full speed, the resulting supernatant was centrifuged at
4°C in an ultracentrifuge (Beckman Coulter) with a 70.1Ti rotor (Beckman
Coulter) at 35,000 rpm for 1 h. 3? SDS sample buffer and DTT (New
England Biolabs) was added to the resulting supernatant, and the samples
were heated to 95°C for 5 min.
For immunoblotting MEFs, cells were lysed by adding Nonidet P-40
lysis buffer (1% Nonidet P-40, 300 mM NaCl, and 50 mM NaH2PO4), with
complete EDTA-free protease inhibitor mixture minitablets added (Roche)
according to the manufacturer’s instructions. After a 5-min incubation on
ice, they were spun in a minicentrifuge at 4°C at full speed for 5 min. 3?
SDS/DTT buffer was added to the resulting supernatant, and the samples
were heated to 95°C for 5 min.
HeLa cell equivalents (1.5 ? 105or 3 ? 105for MEFs) were run on a
12% SDS gel, followed by protein transfer to a nitrocellulose membrane
(Schleicher & Schuell). After transfer, the membrane was rotated overnight
in PBS, 5% milk, and 0.2% Tween 20 to block. After 18 h, the membrane
was stained with rabbit anti-LAP polyclonal Ab (gift from Dr. A. Gold-
berg, Harvard Medical School, Boston, MA) diluted 1/10,000 in PBS,
0.2% Tween 20, and 0.02% NaN3. After 2 h, the blot was washed three
times with PBS and 0.2% Tween 20 for 30 min. The blot was then stained
with HRP-conjugated goat anti-rabbit diluted 1/50,000. The blot was de-
veloped with ECL (Pierce).
Peptide injection and measurement
Peptides were injected into cells, and their half-lives were measured as
previously described (11). The synthesis and sequences of the internally
quenched peptides have been previously described in detail (8). Analysis of
peptide degradation rates was performed as previously described (11).
Exons 1–13 are shown as boxes, with the coding regions in black. The location of the primer sequences used for PCR genotype analysis are shown with
arrows. B, PCR genotype analysis of WT and LAP-deficient animals. Amplification of the WT allele resulted in an 800-bp fragment, whereas the disrupted
allele produced a 200-bp fragment. C, LAP Western blot analysis of cell lysates prepared from MEFs with or without treatment with IFN-?.
Generation of LAP-deficient mice. A, Genomic organization of the mouse LAP gene (upper) and structure of the targeting vector (lower).
6607 The Journal of Immunology
Small interfering RNA transfection of shRNA-stable cells
LAP-KD or shRNA control cells were transfected with siRNA specific for
ERAP1 or with a control siRNA (directed against murine TOP in a region
that differs from human TOP) as previously described (10).
Generation of LAP-deficient mice
LAP-deficient mice were generated using VelociGene (Regeneron
Pharmaceuticals) technology (24). Homologous recombination re-
sulted in the loss of all exons and introns of the LAP genomic
sequence (Fig. 1), beginning 6 bp downstream of the LAP start
codon and ending 624 bp after the LAP stop codon. In total, ?19.2
kb was deleted. The presence of the neogene and subsequent loss of
the entire LAP gene were confirmed by PCR (Fig. 1B). LAP?/?mice
expressed no LAP protein in any of the organs tested, even after IFN
stimulation (Fig. 1C).
Lymphocyte ratios and MHC class I levels in LAP?/?mice are
Because MHC class I surface expression is dependent on peptide
supply, changes in peptide supply can be detected by measuring
surface MHC class I levels. Elimination of LAP could, in princi-
ple, reduce the peptide supply to MHC class I (if LAP predomi-
nately trims peptides longer than nine to 10 aa) or increase peptide
supply (if LAP preferentially trims peptides nine to10 aa in length
that could otherwise bind to MHC class I). We therefore evaluated
the expression of MHC class I on cells in the LAP-deficient ani-
mals. The levels of H-2Dband H-2Kbfrom LAP?/?mice were
similar to those on splenocytes from WT C57BL/6 mice (Fig. 2).
The small reduction in Kbexpression on LAP?/?cells in this
experiment was not statistically significant and was not observed
in other experiments. LAP is normally induced by IFN (15, 30),
and in WT mice, LAP protein levels increased after i.p. injection
of the type I IFN-inducer, poly I:C. However, even under these
conditions, surface levels of H-2Dband H-2Kbwere not signifi-
cantly different in WT and LAP?/?mice (Fig. 2).
Mice lacking important components of the MHC class I Ag
presentation pathway sometimes have reduced levels of CD8?T
lymphocytes due to reduced positive selection in the thymus (31–
33). Therefore, we examined the levels of CD8?T cells in LAP-
deficient mice. Parenchymal, mesenteric, and maxillary lymph
nodes and spleens were harvested from C57BL/6 and LAP?/?
mice, and their CD4?and CD8?T cells were enumerated by flow
cytometry. CD4?B220?cells and CD8?B220?cells were present
mice. FACS analysis of H-2Kb (AF6-88.5; A) and H-2Db (KH95; B) expres-
sion on splenocytes 24 h after i.p. injection with poly I:C or PBS compared
with staining with isotype control. Graphs represent the average geometric
mean fluorescence (GMFI) of two mice (PBS) or three mice (poly I:C) from
each genotype. u, BL6; ?, LAP?/?. Error bars represent the SD within each
group. Data are representative of three independent experiments.
MHC class I presentation in LAP-deficient and C57BL/6
treated with IFN-? (B) were analyzed by flow cytometry after staining for
H-2Kbwith AF6 or Y3 mAbs or for total MHC class I with M1/42 mAb.
Bars represent the average geometric mean fluorescence (GMFI) of four
independent MEF lines for each genotype. u, WT MEFs; ?, LAP-defi-
cient MEFs. Error bars represent the SD within each group.
MHC class I presentation on MEFs. MEFs (A) or MEFs
6608ROLE OF LEUCINE AMINOPEPTIDASE IN Ag PRESENTATION
in similar ratios and numbers in both strains, suggesting that T cell
generation and maturation are grossly normal in LAP?/?mice
(data not shown).
LAP?/?MEFs present peptide as well as WT MEFs
Several independent lines of MEFs were generated from WT and
LAP?/?embryos. As with lymphocytes, no significant differences
in surface H-2Db, H-2Kb, or total MHC class I levels were de-
tected either under normal conditions or after stimulation with
IFN-? (Fig. 3). Using real-time PCR, we detected no increase in
BH, PSA, or ERAP1 expression (data not shown). Therefore, these
other aminopeptidases were not up-regulated and compensating
for a lack of LAP.
Cleavage of full-length OVA by the proteasome in vitro gen-
erates SIINFEKL as well as SIINFEKL extended at the N terminus
by up to 12 residues (3). Similarly, N-extended precursors of SI-
INFEKL are generated from full-length OVA constructs and re-
quire trimming for presentation in vivo (10). To examine the role
of LAP in this process, we transfected WT or LAP?/?MEFs with
plasmids expressing full-length OVA and GFP, so that transfected
cells could be identified by GFP expression. WT and LAP?/?
MEFs (gated for comparable GFP expression) generated similar
amounts of SIINFEKL-Kbcomplexes (Fig. 4A), as detected by
flow cytometry usingthe
IFN-? (B) were analyzed by flow cytometry. FACS traces represent gated cells expressing comparable amounts of GFP. SIINFEKL presentation by H2-Kb
was determined using 25.D1.16. SIINFEKL presentation by two independent LAP-deficient MEFs (LAP KO1 and -2; gray line and black dotted line,
respectively) and WT MEFs (black solid line) were compared with background staining (ctrl; gray filled line) with an isotype control Ab.
SIINFEKL presentation by MEFs. Forty-eight hours after transfection with the indicated peptide construct, MEFs (A) or MEFs treated with
stimulated on day 7 for VSV (A), day 9 for LCMV (B), or day 7 for recombinant Vac (C). After isolation, splenocytes were stimulated for 5 h with
anti-CD3?, as a control for CTL viability; with VSV peptide nuclear protein 52–59 (A), LCMV peptides gp33, np205, gp276, and np396 (B); with Vac
peptide p10 (C); or with SIINFEKL peptides as described in Materials and Methods. They were then surface stained with anti-CD8 and anti-CD44, and
intracellularly stained with anti-IFN-?. Graphs and table represent the average percentages of CD8 T cells that were IFN-? positive (n ? 5 mice). u, BL6;
?, LAP?/?. Error bars represent the SD within each group. There was no significant difference between LAP?/?and WT mice in their response to any
of the epitopes tested.
Intracellular IFN-? staining of peptide-specific T cells. Spleen cells from LAP?/?and WT mice infected with virus were harvested and
6609The Journal of Immunology
To more thoroughly examine the dependence of the processing
of N-extended peptides on aminopeptidases, we expressed in these
MEFs N-extended forms of SIINFEKL from minigenes. In this
situation, all peptides are produced in the cytosol as precursors that
require trimming for presentation (13, 14). SIINFEKL generation
from precursors extended by one (the initiating methionine) or five
residues was identical in WT and LAP?/?MEFs (Fig. 4A). Be-
cause these 9- to 13-residue antigenic precursors can be trans-
ported into the ER and trimmed by ERAP1 (10), we also examined
the presentation of a 33-residue precursor that is too long to be
efficiently transported into the ER by TAP (34). This construct,
which has a 25-residue N-terminal extension, was trimmed and
presented similarly by both WT and LAP?/?MEFs (Fig. 4A).
Therefore, under constitutive conditions, LAP is not essential for
the trimming of N-extended SIINFEKL precursors.
Aminopeptidases can also destroy antigenic peptides by trim-
ming them below the size required for stable binding to MHC class
I molecules. To test whether LAP might be destroying some pre-
sentable peptides, MEFs were transfected with a construct encod-
ing a ubiquitin-SIINFEKL fusion protein. When expressed in cells,
the N-terminal ubiquitin is cleaved, thereby producing the mature
epitope SIINFEKL. If LAP were to trim SIINFEKL, the resulting
products would not bind stably to H-2Kb. However, LAP?/?and
WT cells equivalently presented SIINFEKL generated from the
ubiquitin fusion construct (Fig. 4A). Therefore, LAP does not limit
the amount of SIINFEKL available for presentation.
IFN treatment of cells alters the composition and enzymatic ac-
tivities of the proteasome, resulting in relatively more N-extended
precursors of antigenic epitopes (3) as well as increased levels of
LAP (15). However, even though IFN treatment increased overall
presentation, the presentation of SIINFEKL from each precursor
was the same in WT and LAP?/?MEFs (Fig. 4B), suggesting that
even at induced levels, LAP is not essential in the generation of
some MHC class I peptides.
Presentation of viral epitopes is not altered in LAP?/?mice
To test Ag processing in vivo, we examined the CTL responses to
several viruses. WT and LAP?/?mice were infected with VSV,
LCMV, or a recombinant Vac-OVA. At the peak of each infection,
splenic lymphocytes were isolated, stimulated in vitro with the
appropriate antigenic peptide, and stained for intracellular IFN-?
levels. The frequency of CTL (Fig. 5) and their level of production
of IFN-? (not shown) compared with the VSV peptide (Fig. 5A),
to four antigenic LCMV peptides (Fig. 5B), and to two recombi-
nant Vac-OVA peptides (Fig. 5C) in LAP?/?mice were equal to
those of WT mice, suggesting that neither the quantity nor the
quality of the CTL response in vivo was dependent on LAP.
LAP mRNA expression and protein levels are reduced in shRNA
stable transfectants of HeLa cells
The finding that elimination of LAP does not alter Ag presentation
in mouse cells in vivo or in vitro was unexpected because LAP has
been shown to process N-extended SIINFEKL peptides in cell ly-
sates (15). Because the previous studies used HeLa cell extracts, it
seemed possible that LAP might play a more important role in
human cells or in particular cell types. To investigate this issue,
HeLa-Kbcells (10) were stably transfected with a plasmid encod-
ing a shRNA targeting LAP or encoding a control shRNA that
does not recognize human sequences. Clone LAP-KD showed a
reduction in LAP of 90–95% by real-time PCR (Fig. 6A) and
real-time PCR showed no significant change over time in the expres-
sion of BH or PSA between these cell lines (data not shown).
LAP is inducible by IFN-? in HeLa cells (15). Therefore, we
examined the silencing of LAP in LAP-KD cells treated with
IFN-?. After 48 h of IFN-? treatment, LAP mRNA was increased
?10-fold in the control cell line (Fig. 6A), but LAP mRNA levels
in the LAP-KD cell line, although increased, were still 90–95%
lower than in the control (IFN-stimulated) cell line. Analysis of
protein levels by semiquantitative immunoblotting gave similar
results (Fig. 6B). LAP-KD constitutively had 5–10% as much LAP
as the control cell line. Although treatment with IFN-? increased
the levels of LAP protein ?10-fold in both the LAP-KD and con-
trol cell lines, LAP levels in LAP-KD were ?10-fold lower than
in control cells treated with IFN-?.
Reduction of LAP has no effect on surface presentation of H-
2Kb-SIINFEKL in HeLa cells
Because LAP was originally described as an enzyme in HeLa cell
extracts that trimmed N-terminally extended precursors of the
OVA peptide SIINFEKL, we tested such peptides in LAP-KD
clone stably transfected with a shRNA construct targeting LAP (LAP-KD)
was compared with control cell lines. A, Results from real-time PCR for
LAP mRNA in LAP-KD cells and control cells with or without treatment
with IFN-?, all normalized to ?-actin mRNA. Data are presented as a
percentage of the total LAP mRNA from the control cell line under con-
stitutive conditions (1 ? 100%). Error bars represent variations among
triplicate wells. B, Western blot for LAP of LAP-KD cells and the parental
cell line with or without treatment with IFN-?. Serial dilutions (1/3) of the
HKb?IFN cell lysate were run on the gel to determine relative amounts of
LAP protein between samples (lanes 5–8). According to semiquantitative
immunoblotting and real-time PCR, LAP-KD cells had 90–95% less LAP
mRNA and protein, with or without IFN-? treatment.
LAP expression in the LAP-KD clone. A HeLa-Kbcell
6610ROLE OF LEUCINE AMINOPEPTIDASE IN Ag PRESENTATION
cells. Therefore, LAP-KD cells and control cells were transiently
transfected with plasmids encoding OVA or N-extended SIIN
FEKL precursors, as described above. For each of the constructs,
the generation of H-2Kb-SIINFEKL was the same in control and
LAP-KD cells (Fig. 7A). There was also no difference between the
control and LAP-KD cells when they were treated with IFN-?
(Fig. 7B). Therefore, LAP is not required for generation of MHC
class I peptides in HeLa cells.
N-terminally extended peptides may be transported to the ER by
TAP, where they can be trimmed to mature epitopes by ERAP1.
To determine whether ERAP1 was masking an effect of LAP de-
ficiency, we examined the phenotype of cells in which both LAP
and ERAP1 were silenced. LAP-KD and control cells were treated
with an siRNA targeting ERAP1 or with control siRNA (10) and
then transiently transfected with plasmids encoding N-terminally
extended SIINFEKL precursors, as described above. As previously
reported (10), presentation of SIINFEKL from N-extended precur-
sors was reduced in HeLa cells treated with ERAP1 siRNA (Fig.
8). In comparison, SIINFEKL presentation was not further reduced
in LAP and ERAP1 double-KD cells. Presentation of the SIIN
FEKL precursor was also not different in ERAP1-deficient vs
LAP- and ERAP1-deficient cells after stimulation with IFN-? for
24 h after transfection (Fig. 8).
cells (A) or LAP-KD and control cells treated with IFN-? (B) were analyzed by flow cytometry. FACS traces represent gated cells expressing comparable
amounts of GFP. SIINFEKL presentation by H2-Kb was determined using 25.D1.16. SIINFEKL presentation by LAP-KD cells (gray line) and control cells
(black dotted line) were compared with background staining (gray-filled line) using an isotype control Ab.
SIINFEKL presentation by LAP-KD cells. Forty-eight hours after transfection with the indicated peptide construct, LAP-KD and control
ERAP1 or control (ctrl) siRNA. Twenty-four hours later (day 1), all cells were transfected with N25?SIINFEKL as previously described. On day 3, all
cells were analyzed by FACS. For the cells in B, IFN-? was added on day 2. FACS traces represent gated cells expressing comparable amounts of GFP.
SIINFEKL presentation by H-2Kbwas determined using 25.D1.16. SIINFEKL presentations by LAP-KD?ctrl siRNA (gray thin line), LAP-KD?ERAP1
siRNA (black solid line), ctrl cells?ctrl siRNA (black dashed line), and ctrl cells?ERAP1 siRNA (gray thick line) were all compared with those by cells
transfected with vector and stained with 25.D1.16 (gray-filled line). Although reduction of ERAP1 reduced presentation of SIINFEKL, this reduction was
not enhanced or decreased by the loss of LAP either under constitutive conditions (A) or after 24-h incubation with IFN-? (B).
SIINFEKL presentation on shRNA-stable cells treated with ERAP1 siRNA. LAP-KD or control cells were treated with siRNA targeting
6611The Journal of Immunology
Reduction of LAP has no effect on the rate of peptide trimming
The quantitation of Ag presentation is an indirect measure of the
generation of antigenic peptides. It is possible that a contribution
of LAP to this process could be missed if the trimming of peptides
was not a rate-limiting step in the pathway. We therefore directly
measured the rate of peptide degradation in control and LAP-KD
cells by microinjection of peptides containing fluorescein and
quencher adducts (8, 11). These substrates generate a fluorescent
signal when aminopeptidases cleave and thereby separate the res-
idues containing the fluorophore and quencher moieties. Peptides
with different amino acids in the P1 position were microinjected
into cells, and fluorescence resulting from trimming the peptides
Although there was some variation in the rate of trimming of
different peptide sequences, there was no statistically significant
difference in the half-life of any of the tested peptides between the
two cell lines (Fig. 9A). Treating the cells with IFN-? before mi-
croinjection increased LAP expression (Fig. 9B), but did not alter
the rate of processing of the peptides (Fig. 9, A and C). These data
demonstrate that LAP activity does not influence the rate of pep-
tide trimming in the cytosol.
In this study, we show 1) that LAP is not essential for viability or
normal development of mice, and 2) that LAP does not play an
indispensable role in generating peptides presented by MHC class
I under constitutive conditions or after stimulation with IFN. These
results indicate that there is considerable functional redundancy
among aminopeptidases within cells.
There is considerable evidence that aminopeptidases are impor-
tant in trimming antigenic precursors for presentation. Protea-
somes have been shown in vitro (3, 35) and in vivo (10) to generate
many N-extended precursors. Other experiments have shown that
if such N-extended peptide precursors are expressed or injected
into cells, they are trimmed and presented by MHC class I on the
cell surface (11, 14). Processing of such peptides can occur in the
presence of proteasome inhibitors (14, 36, 37), indicating that
other peptidases in the cell can cleave N-terminal residues. Amin-
opeptidases can trim N-extended precursors that have a free N
terminus, but not if the N terminus is blocked (11, 13). Finally,
eliminating or blocking aminopeptidases in vivo reduces Ag pre-
sentation (10, 17, 38). Together, these data suggest that N-ex-
tended precursors are generated during protein degradation and
then trimmed by aminopeptidases in vivo.
N-extended peptides that are preceded by an ER-localizing sig-
nal sequence are efficiently trimmed to mature epitopes in TAP-
deficient cells, demonstrating that aminopeptidases localized in the
ER can generate peptides for presentation (14, 39, 40). In cells
lacking ERAP1, N-extended peptides targeted to the ER are not
presented (9, 10) indicating that ERAP1 plays an essential, non-
redundant role in trimming in the ER. However, several lines of
evidence indicate that N-extended peptides can also be trimmed in
the cytosol before their transport into the ER. First, cytosolic N-
extended SIINFEKL precursors expressed in ERAP1-deficient
cells are trimmed and presented (albeit in reduced amounts),
whereas when the same constructs are targeted to the ER, SIIN-
FEKL is not presented (10). Second, peptides preceded by N-ex-
tensions of up to 25 aa are too long to be efficiently transported by
TAP, but when expressed in the cytosol, they are still trimmed and
presented by MHC class I in the absence of proteasome activity
(14). Third, cytosolic extracts devoid of ER contaminants can trim
N-extended peptides (15, 17). Fourth, treatment of cells with in-
hibitors of cytosolic aminopeptidases reduces the presentation of
or control cells. The half-life of the peptide was determined by following the generation of fluorescence signal. B, LAP Western blot of LAP-KD and control
cells after incubation with IFN-?; C, peptide half-life, measured in seconds of a microinjected fluorescent peptide with leucine in the P1 position, with or
without incubation with IFN-?. In all experiments, no difference in half-life was detected in any of the tested peptides between the two cell lines.
Half-life of microinjected peptides in LAP-KD cells. A, Peptides with different residues in the P1 position were microinjected into LAP-KD
6612 ROLE OF LEUCINE AMINOPEPTIDASE IN Ag PRESENTATION
some peptides (8, 17). A major unresolved issue in the field is the
identity and extent of contribution of the cytosolic aminopeptidases
that participate in the generation of MHC class I-presented peptides.
Tripeptidyl peptidase II is a cytosolic tripeptidyl peptidase that
has been shown to play an important role in trimming peptides that
are longer than ?16 residues (8). However, it remains unclear
which cytosolic aminopeptidase(s) trims shorter peptides of 15 aa
or less. It has also been unclear how much redundancy exists
among the cytosolic aminopeptidases.
LAP is a cytosolic aminopeptidase that was thought to be a
major contributor to the MHC class I peptide pool for a number of
reasons. First, it was shown to be a major peptidase in cytosolic
extracts that generated SIINFEKL from N-extended precursors.
Second, transcription of LAP is inducible by both type I and type
II IFNs. Because IFNs enhance MHC class I presentation and up-
regulate most components of this pathway, the IFN induction of
LAP suggested that it might have a particularly important role
during inflammation. Third, overexpression of LAP results in a
shorter half-life of peptides in living cells (11). Given these find-
ings, our present results are very surprising.
We have found no detectable defect in the ability of LAP-defi-
cient mice to generate or present MHC class I peptides, even after
treatment with the type I IFN inducer poly I:C. In these experi-
ments we have measured the presentation of a number of peptides
to CTLs, including the antigenic model epitope SIINFEKL, which
is a well-characterized substrate of LAP. No qualitative or quan-
titative difference was detected in the priming of CTL responses to
any of these epitopes in LAP-deficient and WT mice. In addition,
we have shown in MEFs and HeLa cells that LAP was not required
for the trimming of peptide precursors from various N-extended
SIINFEKL constructs, even in the presence of IFN-?.
One possibility is that in the absence of LAP, peptides that
would normally be trimmed into the cytosol are transported into
the ER (either in their original form or as an intermediate precursor
that has been partially trimmed by other cytosolic aminopepti-
dases). ERAP in the ER can rapidly process N-extended SIIN
FEKL precursors to SIINFEKL (10), and it was possible that even
if LAP deficiency led to an increased amount of N-extended pre-
cursors entering the ER, that ERAP1 activity could mask this ef-
fect. However, although reducing ERAP1 dramatically reduced the
presentation of SIINFEKL generated from these peptides, there
was no further reduction seen when LAP was also knocked down
(Fig. 8). Therefore, ERAP1 is not masking a contribution of LAP
to trimming precursor peptides.
Although we examined a number of different peptides in this
study, it is still possible that a subset of peptides exists for which
LAP is required for presentation. LAP has been shown to hydro-
lyze different amino acid residues (from amino acid-7-amino-4-
methylcoumarin substrates) at different rates, so LAP may poten-
tially play a more dominant role in trimming certain sequences.
However, when peptides with different N-terminal residues were
injected into LAP-deficient HeLa cells, no difference in peptide
half-life was detected even after stimulation with IFN-?. More-
over, LAP-deficient mouse cells express normal levels of MHC
class I, indicating that the overall supply of peptides to MHC class
I is not changed. Therefore, if there are LAP-dependent epitopes,
they must constitute a minority of peptides.
Together, these findings demonstrate that LAP does not play an
essential role in producing the majority of peptides for MHC class
I presentation. The fact that SIINFEKL was presented from N-
extended precursors, even in LAP and ERAP1 double-deficient
cells indicates that there are other cytosolic aminopeptidases that
can substitute for LAP. Although it is possible that other amin-
opeptidases could be up-regulated to compensate for the absence
of LAP, we have not found any change in BH, PSA, and ERAP1
expression that is correlated with the loss of LAP, in MEFs or of
BH or PSA in LAP-KD cells (data not shown). In any case, our
data indicate that there must be considerable redundancy in the
trimming function of cytosolic aminopeptidases in vivo.
It is possible that the contribution of LAP to peptide supply is
not apparent because the trimming step is not rate limiting. This
would explain why peptides with one, two, or three extra residues
on the N terminus are presented with similar kinetics to one an-
other and to the mature epitope itself, at least in some systems (14)
(data not shown). Under these conditions, other aminopeptidases
would be sufficient to trim antigenic precursors to mature epitopes.
Surprisingly, we also found no evidence that LAP destroys ma-
ture epitopes, thereby limiting their presentation. There was no
increase in peptide supply in the absence of LAP. It is possible that
this is because LAP destroys as many peptides as it produces;
however, our data argue against such a possibility. First, LAP-
deficient cells show no defect in generating and presenting the
mature SIINFEKL epitope from the ubiquitin-SIINFEKL fusion
construct, even though trimming of just one residue would gener-
ate IINFEKL, which is too short to bind H-2Kb(Fig. 4). Second,
the hydrolysis rates of individually injected peptides are not dif-
ferent in LAP-KD cells. For these reasons, we believe that LAP
does not play a major role in the destruction of antigenic peptides
under physiological conditions. However, it should be noted that
when LAP is expressed at supraphysiological levels, peptides are
hydrolyzed more rapidly in cells (11).
The redundancy of aminopeptidases is likely to be useful to the
cell in some way. Redundancy would ensure optimal recycling of
peptides into amino acids, which would conserve energy for the
cell. It would also prevent the build-up of proteasomal products,
which could interfere with protein-protein interactions or be toxic
in other ways. Perhaps it is for these reasons that LAP-deficient
mice are viable and demonstrate a normal phenotype.
However the question remains: if LAP is not essential for Ag
presentation then why is it inducible by IFN? It is known that IFN
stimulation up-regulates many components of the Ag-processing
pathway, including the immunoproteasome. Immunoproteasomes
can generate distinct MHC class I-presented peptides that are not
produced by the constitutive proteasome (41–44) In addition, the
immunoproteasome has been shown to generate longer peptides
than the constitutive proteasome for at least one Ag (3), and there
are indirect data suggesting that this might be generally true (10).
It is possible that in some cell types the immunoproteasome gen-
erates antigenic precursors that are more efficiently processed by
IFN-inducible aminopeptidases such as LAP. Alternatively, it is
possible that LAP is more essential in a pathway not involved in
Ag presentation, such as the breakdown of some other component
involved in an immune response, or in the breakdown of viral
peptides during infection. The LAP-deficient mice and cells should
be useful in future studies to help answer these questions. In ad-
dition, further investigations into the roles of other cytosolic amin-
opeptidases will help to define the redundancy and/or unique roles
of these various peptidases compared with those of LAP.
We thank Dr. Raymond Welsh for providing viruses and peptides.
The authors have no financial conflict of interest.
1. Rock, K. L., I. A. York, and A. L. Goldberg. 2004. Post-proteasomal antigen
processing for major histocompatibility complex class I presentation. Nat. Im-
munol. 5: 670–677.
6613 The Journal of Immunology
2. Rock, K. L., C. Gramm, L. Rothstein, K. Clark, R. Stein, L. Dick, D. Hwang, and Download full-text
A. L. Goldberg. 1994. Inhibitors of the proteasome block the degradation of most
cell proteins and the generation of peptides presented on MHC class I molecules.
Cell 78: 761–771.
3. Cascio, P., C. Hilton, A. F. Kisselev, K. L. Rock, and A. L. Goldberg. 2001. 26S
proteasomes and immunoproteasomes produce mainly N-extended versions of an
antigenic peptide. EMBO J. 20: 2357–2366.
4. Kisselev, A. F., T. N. Akopian, and A. L. Goldberg. 1998. Range of sizes of
peptide products generated during degradation of different proteins by archaeal
proteasomes. J. Biol. Chem. 273: 1982–1989.
5. Emmerich, N. P., A. K. Nussbaum, S. Stevanovic, M. Priemer, R. E. Toes,
H. G. Rammensee, and H. Schild. 2000. The human 26 S and 20 S proteasomes
generate overlapping but different sets of peptide fragments from a model protein
substrate. J. Biol. Chem. 275: 21140–21148.
6. Goldberg, A. L., P. Cascio, T. Saric, and K. L. Rock. 2002. The importance of the
proteasome and subsequent proteolytic steps in the generation of antigenic pep-
tides. Mol. Immunol. 39: 147–164.
7. Saric, T., S. C. Chang, A. Hattori, I. A. York, S. Markant, K. L. Rock,
M. Tsujimoto, and A. L. Goldberg. 2002. An IFN-?-induced aminopeptidase in
the ER, ERAP1, trims precursors to MHC class I-presented peptides. Nat. Im-
munol. 3: 1169–1176.
8. Reits, E., J. Neijssen, C. Herberts, W. Benckhuijsen, L. Janssen, J. W. Drijfhout,
and J. Neefjes. 2004. A major role for TPPII in trimming proteasomal degrada-
tion products for MHC class I antigen presentation. Immunity 20: 495–506.
9. Serwold, T., F. Gonzalez, J. Kim, R. Jacob, and N. Shastri. 2002. ERAAP cus-
tomizes peptides for MHC class I molecules in the endoplasmic reticulum. Na-
ture 419: 480–483.
10. York, I. A., S. C. Chang, T. Saric, J. A. Keys, J. M. Favreau, A. L. Goldberg, and
K. L. Rock. 2002. The ER aminopeptidase ERAP1 enhances or limits antigen
presentation by trimming epitopes to 8–9 residues. Nat. Immunol. 3: 1177–1184.
11. Reits, E., A. Griekspoor, J. Neijssen, T. Groothuis, K. Jalink, P. van Veelen,
H. Janssen, J. Calafat, J. W. Drijfhout, and J. Neefjes. 2003. Peptide diffusion,
protection, and degradation in nuclear and cytoplasmic compartments before an-
tigen presentation by MHC class I. Immunity 18: 97–108.
12. Yewdell, J. W., E. Reits, and J. Neefjes. 2003. Making sense of mass destruction:
quantitating MHC class I antigen presentation. Nat. Rev. Immunol. 3: 952–961.
13. Mo, X. Y., P. Cascio, K. Lemerise, A. L. Goldberg, and K. Rock. 1999. Distinct
proteolytic processes generate the C and N termini of MHC class I-binding pep-
tides. J. Immunol. 163: 5851–5859.
14. Craiu, A., T. Akopian, A. Goldberg, and K. L. Rock. 1997. Two distinct proteo-
lytic processes in the generation of a major histocompatibility complex class
I-presented peptide. Proc. Natl. Acad. Sci. USA 94: 10850–10855.
15. Beninga, J., K. L. Rock, and A. L. Goldberg. 1998. Interferon-? can stimulate
post-proteasomal trimming of the N terminus of an antigenic peptide by inducing
leucine aminopeptidase. J. Biol. Chem. 273: 18734–18742.
16. Serna, A., M. C. Ramirez, A. Soukhanova, and L. J. Sigal. 2003. Cutting edge:
efficient MHC class I cross-presentation during early vaccinia infection requires
the transfer of proteasomal intermediates between antigen donor and presenting
cells. J. Immunol. 171: 5668–5672.
17. Stoltze, L., M. Schirle, G. Schwarz, C. Schroter, M. W. Thompson, L. B. Hersh,
H. Kalbacher, S. Stevanovic, H. G. Rammensee, and H. Schild. 2000. Two new
proteases in the MHC class I processing pathway. Nat. Immunol. 1: 413–418.
18. York, I. A., A. X. Mo, K. Lemerise, W. Zeng, Y. Shen, C. R. Abraham, T. Saric,
A. L. Goldberg, and K. L. Rock. 2003. The cytosolic endopeptidase, thimet
oligopeptidase, destroys antigenic peptides and limits the extent of MHC class I
antigen presentation. Immunity 18: 429–440.
19. Saric, T., J. Beninga, C. I. Graef, T. N. Akopian, K. L. Rock, and A. L. Goldberg.
2001. Major histocompatibility complex class I-presented antigenic peptides are
degraded in cytosolic extracts primarily by thimet oligopeptidase. J. Biol. Chem.
20. Saric, T., C. I. Graef, and A. L. Goldberg. 2004. Pathway for degradation of
peptides generated by proteasomes: a key role for thimet oligopeptidase and other
metallopeptidases. J. Biol. Chem. 279: 46723–46732.
21. Paddison, P. J., A. A. Caudy, E. Bernstein, G. J. Hannon, and D. S. Conklin.
2002. Short hairpin RNAs (shRNAs) induce sequence-specific silencing in mam-
malian cells. Genes Dev. 16: 948–958.
22. Miller, A. T., and L. J. Berg. 2002. Defective Fas ligand expression and activa-
tion-induced cell death in the absence of IL-2-inducible T cell kinase. J. Immunol.
23. Bachmair, A., D. Finley, and A. Varshavsky. 1986. In vivo half-life of a protein
is a function of its amino-terminal residue. Science 234: 179–186.
24. Valenzuela, D. M., A. J. Murphy, D. Frendewey, N. W. Gale, A. N. Economides,
W. Auerbach, W. T. Poueymirou, N. C. Adams, J. Rojas, J. Yasenchak, et al.
2003. High-throughput engineering of the mouse genome coupled with high-
resolution expression analysis. Nat. Biotechnol. 21: 652–659.
25. Restifo, N. P., I. Bacik, K. R. Irvine, J. W. Yewdell, B. J. McCabe,
R. W. Anderson, L. C. Eisenlohr, S. A. Rosenberg, and J. R. Bennink. 1995.
Antigen processing in vivo and the elicitation of primary CTL responses. J. Im-
munol. 154: 4414–4422.
26. Porgador, A., J. W. Yewdell, Y. Deng, J. R. Bennink, and R. N. Germain. 1997.
Localization, quantitation, and in situ detection of specific peptide-MHC class I
complexes using a monoclonal antibody. Immunity 6: 715–726.
27. Kuhns, S. T., and L. R. Pease. 1998. A region of conformational variability
outside the peptide-binding site of a class I MHC molecule. J. Immunol. 161:
28. Albert, F., C. Boyer, L. D. Leserman, and A. M. Schmitt-Verhulst. 1983. Immu-
nopurification and insertion into liposomes of native and mutant H-2Kb: quan-
tification by solid phase radioimmunoassay. Mol. Immunol. 20: 655–667.
29. Stallcup, K. C., T. A. Springer, and M. F. Mescher. 1981. Characterization of an
anti-H-2 monoclonal antibody and its use in large-scale antigen purification.
J. Immunol. 127: 923–930.
30. Craven, R. A., A. J. Stanley, S. Hanrahan, N. Totty, D. P. Jackson, R. Popescu,
A. Taylor, J. Frey, P. J. Selby, P. M. Patel, et al. 2004. Identification of proteins
regulated by interferon-? in resistant and sensitive malignant melanoma cell
lines. Proteomics 4: 3998–4009.
31. Van Kaer, L., P. G. Ashton-Rickardt, H. L. Ploegh, and S. Tonegawa. 1992.
TAP1 mutant mice are deficient in antigen presentation, surface class I molecules,
and CD4?8?T cells. Cell 71: 1205–1214.
32. Van Kaer, L., P. G. Ashton-Rickardt, M. Eichelberger, M. Gaczynska,
K. Nagashima, K. L. Rock, A. L. Goldberg, P. C. Doherty, and S. Tonegawa.
1994. Altered peptidase and viral-specific T cell response in LMP2 mutant mice.
Immunity 1: 533–541.
33. Zijlstra, M., M. Bix, N. E. Simister, J. M. Loring, D. H. Raulet, and R. Jaenisch.
1990. ?2-Microglobulin deficient mice lack CD4?8?cytolytic T cells. Nature
34. Koopmann, J. O., M. Post, J. J. Neefjes, G. J. Hammerling, and F. Momburg.
1996. Translocation of long peptides by transporters associated with antigen pro-
cessing (TAP). Eur. J. Immunol. 26: 1720–1728.
35. Komlosh, A., F. Momburg, T. Weinschenk, N. Emmerich, H. Schild, E. Nadav,
I. Shaked, and Y. Reiss. 2001. A role for a novel luminal endoplasmic reticulum
aminopeptidase in final trimming of 26 S proteasome-generated major histocom-
patability complex class I antigenic peptides. J. Biol. Chem. 276: 30050–30056.
36. Golovina, T. N., E. J. Wherry, T. N. Bullock, and L. C. Eisenlohr. 2002. Efficient
and qualitatively distinct MHC class I-restricted presentation of antigen targeted
to the endoplasmic reticulum. J. Immunol. 168: 2667–2675.
37. Stoltze, L., T. P. Dick, M. Deeg, B. Pommerl, H. G. Rammensee, and H. Schild.
1998. Generation of the vesicular stomatitis virus nucleoprotein cytotoxic T lym-
phocyte epitope requires proteasome-dependent and -independent proteolytic ac-
tivities. Eur. J. Immunol. 28: 4029–4036.
38. Levy, F., L. Burri, S. Morel, A. L. Peitrequin, N. Levy, A. Bachi, U. Hellman,
B. J. Van den Eynde, and C. Servis. 2002. The final N-terminal trimming of a
subaminoterminal proline-containing HLA class I-restricted antigenic peptide in
the cytosol is mediated by two peptidases. J. Immunol. 169: 4161–4171.
39. Snyder, H. L., J. W. Yewdell, and J. R. Bennink. 1994. Trimming of antigenic
peptides in an early secretory compartment. J. Exp. Med. 180: 2389–2394.
40. Snyder, H. L., I. Bacik, J. W. Yewdell, T. W. Behrens, and J. R. Bennink. 1998. Pro-
miscuous liberation of MHC-class I-binding peptides from the C termini of membrane
and soluble proteins in the secretory pathway. Eur. J. Immunol. 28: 1339–1346.
41. Gileadi, U., H. T. Moins-Teisserenc, I. Correa, B. L. Booth, Jr., P. R. Dunbar,
A. K. Sewell, J. Trowsdale, R. E. Phillips, and V. Cerundolo. 1999. Generation
of an immunodominant CTL epitope is affected by proteasome subunit compo-
sition and stability of the antigenic protein. J. Immunol. 163: 6045–6052.
42. Schultz, E. S., J. Chapiro, C. Lurquin, S. Claverol, O. Burlet-Schiltz, G. Warnier,
V. Russo, S. Morel, F. Levy, T. Boon, et al. 2002. The production of a new
MAGE-3 peptide presented to cytolytic T lymphocytes by HLA-B40 requires the
immunoproteasome. J. Exp. Med. 195: 391–399.
43. Schwarz, K., M. van Den Broek, S. Kostka, R. Kraft, A. Soza, G. Schmidtke,
P. M. Kloetzel, and M. Groettrup. 2000. Overexpression of the proteasome sub-
units LMP2, LMP7, and MECL-1, but not PA28 ?/?, enhances the presentation
of an immunodominant lymphocytic choriomeningitis virus T cell epitope. J. Im-
munol. 165: 768–778.
44. Sijts, A. J., S. Standera, R. E. Toes, T. Ruppert, N. J. Beekman, P. A. van Veelen,
F. A. Ossendorp, C. J. Melief, and P. M. Kloetzel. 2000. MHC class I antigen
processing of an adenovirus CTL epitope is linked to the levels of immunopro-
teasomes in infected cells. J. Immunol. 164: 4500–4506.
6614 ROLE OF LEUCINE AMINOPEPTIDASE IN Ag PRESENTATION